改进小波多分辨力分析的数据融合算法  

Data fusion algorithm based on modified wavelet multi-resolution analysis

作　　者：刘素一[1] 张海霞[1] 罗维平[1] 

机构地区：[1]武汉科技学院电信系,湖北武汉430073

出　　处：《传感器与微系统》2007年第8期105-107,共3页Transducer and Microsystem Technologies

基　　金：湖北省教育厅科研计划资助项目(D200517001)

摘　　要：在小波分析中,多分辨力分析是一重要的方法。但该方法只对低频段的近似信号进行逐级细分,导致低频段频率分辨力越来越高,而高频段的细节信号保持不变,频率分辨力较低。为了克服小波多分辨力分析在高频段频率分辨力低的缺点,采用改进的小波多分辨力分析方法,由于该方法对高频段进行逐级细分,改善了小波变换在高频段的时频局部化性能,提高了小波变换高频段频率的分辨力。同时,将改进的小波多分辨力分析方法应用到多传感器数据融合中,经过仿真计算,其结果表明:该算法是有效的。Multi-resolution analysis is an important method of wavelet analysis. It shows the shortcoming of low resolution in high frequency when the method is used to subdivide the approximate signal in low frequency step by step ,which makes the resolution of low frequency higher and higher and maintains the detail signal of high frequency. To overcome the wavelet multi-resolution analysis shortcoming of low resolution in high frequency bands,a modified wavelet multi-resolution analysis algorithm is proposed, The modified method can subdivide the signal in high frequency bands step by step, which can improve time-frequency characteristic of wavelet transform and increase resolution of high frequency bands. Moreover, the modified wavelet multi-resolution analysis is applied into multi-sensor data fusion, and simulation experiment verifies the validity of this method.

关 键 词：多分辨力分析 改进的多分辨力分析 卡尔曼滤波器 传感器 数据融合 

分 类 号：TN911.72[电子电信—通信与信息系统]